236 
ASTRONOMY: F. H. SEARES 
found in the fact that these A stars, whose absolute magnitudes, with one 
exception, he within the range If = +1.2 to +3.0, agree with the curve for 
M = 0 determined by the adjacent spectral types (see fig. 1,) only after 
their colors have been reduced to M = 0 in accordance with the relation 
shown in fig. 2. 
The giant and dwarf F stars in the mean have sensibly the same color, but 
when the individual results are plotted we find pretty definite evidence of a 
change in color as indicated in fig. 2, first an increase and then a decrease 
in color with increasing M, the maximum lying somewhere near M = + 2. 
The results for the G's and K's (fig. 3;) are quahtatively certain beyond 
any question, the giants as before remarked being very much redder than 
the dwarfs. For the K stars the difference in the corresponding color indicates 
for M = —1 and +7 is more than 0.6 magnitude. This large difference is 
-4 
8.9 
8.8 
9.2 
9.1 
-2 
• 
* • 
•* • 
: B 
5 
• 
M 
As 
• 
• * • 
• • 
• 
• 
F 
• 
9.1 
9.0 
FIG. 2. VARIATION OF COLOR (LOGARITHM OF EXPOSURE-RATIO, AS ORDINATE) WITH 
ABSOLUTE MAGNITUDE (ABSCISSA) FOR B, A AND F STARS 
The points represent individual stars whose colors have been . reduced by the curves of 
Fig. 1 to the mean spectra, B5, A5, etc., for the respective classes. The crosses in the 
A5 group are FO and Fl stars reduced to A5, 
immediately suggestive of the errors that may occur in attempting to infer 
spectral type from color, or to determine visual magnitude from photographic 
with the aid of the spectral type, as is not infrequently done. 
Among the dwarfs themselves, neither for the G's nor the K's, does there 
appear to be much change of color, although the faintest K's suggest that a 
minimum occurs at M = +7 and that from there on the color increases with 
increasing M. 
The last curve of fig. 3 illustrates the similarity of color for the giant and 
dwarf M stars. 
In spite of the diversity of the curves, relationships for adjacent spectral 
classes can usually be traced. Thus the curve for the A stars seems to be re- 
lated to the ascending branch of that for the F's, while the descending branch 
of the latter curve reappears in the results for the G's and K's. The behavior 
